Yes, it is correct.
I slightly modified your circuit.
No capacitors, except for tank caps.
No phase shift.
Quite clear, 11Vp-p sine:
View attachment 186415

EDIT: @MisterBill2

View attachment 186423

This is amazing!
Falstad link: https://bit.ly/2m2KGap

The thing that caught my attention was the R2 - 2K resistor.

I assume that the R2 base resistor prevents excess current from flowing into the base, when the transistor is near saturation.
If its not there, current gets distorted.

Is that correct ?

 

I assume that the R2 base capacitor prevents excess current from flowing into the base, when the transistor is near saturation.
If its not there, current gets distorted.

R2 is resistor, not capacitor.
Resistors R1 and R2 form voltage divider for transistor biasing,
like resistors 22k and 10k in circuit shown in post #1.

I guess your initial question about distortion is fully answered.

BTW, use OUT_2 if you need signal with very low distortion:

 

R2 is resistor, not capacitor.
Resistors R1 and R2 form voltage divider for transistor biasing,
like resistors 22k and 10k in circuit shown in post #1.

I guess your initial question about distortion is fully answered.

BTW, use OUT_2 if you need signal with very low distortion:
View attachment 186555

sorry for the mistake, i meant resistor.

yeah, i now understand the distortions a lot better and learned many things about oscillators.
Transistors should be kept in linear zone (away from cutoff, saturation), closed loop phase shift to be kept at minimum.

I am also following the oscillator showed in post#4 and if i have any questions about it, shall post in a new thread.

Thanks for the massive help.

 

Based on the harmonic content shown in post 19, just because I don't see the distortion does not mean it is not there. BUT certainly there is no obvious tendency to break into a higher order oscillation either.So the circuit is an improvement. Probably changing R1 to 100K, R 2 to 27K, and R4 to 1K will reduce the harmonics quite a bit.

hi
would it be correct to say that

more distortions = more harmonics = poor frequency stability

i mean, do they all indicate the same thing ?

 

hi
would it be correct to say that

more distortions = more harmonics = poor frequency stability

i mean, do they all indicate the same thing ?

No, they do not mean the same thing at all.Frequency stability is entirely different from harmonic content and waveform distortion. Under some conditions there might be the same condition causing both, but a wandering frequency can have a very good sine waveform, while a square wave could be very steady.
Distortion is caused by non-linear operation at some point in the circuit, while frequency change is usually due to changes in some change in the frequency determining parameters, either reactances or voltages.
The tendency to have a second oscillation frequency at some parts of the intended frequency waveform is the result of an instability of the circuit leading to another feedback path at some point in the cycle, as non-linear operation increases the gain at another frequency. That is the cause of that distortion of the waveform that is seen in post #1.

 

Origin of harmonics:

 

The waveform in post 26 is a very good explanation of what happens when an oscillator is set up for rapid starting instead of considering minimum distortion. In the example, adding a series emitter resistor of about 1K ohm, bypassed by a capacitor of 0.01Mfd will provide a short period of higher gain for a quick start and then reduce the gain so that the distortion is reduced a lot. Possibly even a much smaller value of capacitor will also work.

 

@silv3r.m00n, @MisterBill2:

Simple circuit in post #22 works in controlled clipping mode,
so it has amplitude stabilization.
Long term stability of sine amplitude depends on stability
of power supply voltage.
Dependence amplitude on temperature also is compensated.
In temperature range -10°C ... +70°C amplitude changes on 0.66%.
Due to π-filter C2, L1, C1 upper harmonics are attenuated,
and sine signal on OUT_2 has distortion factor (THD) k = 0.176%,
or distortion attenuation a = -55.1dB.

EDIT: @silv3r.m00n - links
https://sound-au.com/articles/sinewave.htm

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.452.1047&rep=rep1&type=pdf

http://rfic.eecs.berkeley.edu/142/pdf/module21.pdf

http://164.100.133.129:81/eCONTENT/Uploads/Session 6.pdf

https://www.edn.com/design/analog/4323726/LC-oscillator-has-stable-amplitude

http://www.circuitstoday.com/lc-oscillators-and-types

https://labcit.ligo.caltech.edu/~vsanni/ph5/pdf/Ph5.Chapter.BasicsOnOscillators.20121030.1244.pdf

https://www.audio-perfection.com/spice-ltspice/distortion-measurements-with-ltspice.html

Circuit with oscillations in linear area:
Distortion factor k = 0.087% at 25°C (OUT_2)

 

@silv3r.m00n, @MisterBill2:

Simple circuit in post #22 works in controlled clipping mode,
so it has amplitude stabilization.
Long term stability of sine amplitude depends on stability
of power supply voltage.
Dependence amplitude on temperature also is compensated.
In temperature range -10°C ... +70°C amplitude changes on 0.66%.
Due to π-filter C2, L1, C1 upper harmonics are attenuated,
and sine signal on OUT_2 has distortion factor (THD) k = 0.176%,
or distortion attenuation a = -55.1dB.

that's some massive insight. i am wondering about a few things:

1. what is "controlled clipping mode" exactly ?
2. what disturbances does the power supply suffer over time, besides lowering of voltage ?

 

With oscillators there are also concerns about fast starting and frequency stability. Of course increasing the feedback to the point of running the active device into it's non-linear areas is always going to produce distortion. That is inescapable. An optimum design includes negative feedback that tends to keep the oscillation in the linear area of the active device, presuming that the intention is to produce a low distortion sine wave. Oscillators intended to produce strong harmonics are a different case.

i am looking for some circuit ideas on "negative feedback to keep oscillations in linear area ..."
could you share some link or tutorial ?

thanks

 

A very simple way to add some negative amplitude feedback is either an emitter resistor or a source resistor, since as the current increases it tends to reduce the turn-on bias and the gain. Setting the correct value may require a bit of math, and possibly even some experimenting. That is why often there are emitter resistors used in bipolar transistor circuits, and even in vacuum tube circuits. A bit of math is needed but it is not hard.

 

@silv3r.m00n, @MisterBill2:

Simple circuit in post #22 works in controlled clipping mode,
so it has amplitude stabilization.
Long term stability of sine amplitude depends on stability
of power supply voltage.
Dependence amplitude on temperature also is compensated.
In temperature range -10°C ... +70°C amplitude changes on 0.66%.
Due to π-filter C2, L1, C1 upper harmonics are attenuated,
and sine signal on OUT_2 has distortion factor (THD) k = 0.176%,
or distortion attenuation a = -55.1dB.
...
Circuit with oscillations in linear area:
Distortion factor k = 0.087% at 25°C (OUT_2)
View attachment 186781

I saw a similar thing in ltspice when playing with the circuit. Increasing the emitter resistance to around 137 ohms reduces gain,
and eliminates much of the distortion.

I guess the loop gain has to be close to 1 ( or exactly 1 in theory ) for a pure distortion free sine wave.
Gain slightly lower than 1 would lead to decay, and over 1 would lead to amplitude clipping and eventually distortion.

There happens to be an old thread explaining the same thing -
Distortion in an Oscillator

This needs component values to be chosen carefully.
I wonder if there can be an automatic gain control mechanism that reduces gain when transistor cross the limits of linear operation ?

Thanks for the links, thats a gold mine of information.

 

I saw a similar thing in ltspice when playing with the circuit. Increasing the emitter resistance to around 137 ohms reduces gain,
and eliminates much of the distortion.

I guess the loop gain has to be close to 1 ( or exactly 1 in theory ) for a pure distortion free sine wave.
Gain slightly lower than 1 would lead to decay, and over 1 would lead to amplitude clipping and eventually distortion.

There happens to be an old thread explaining the same thing -
Distortion in an Oscillator

This needs component values to be chosen carefully.
I wonder if there can be an automatic gain control mechanism that reduces gain when transistor cross the limits of linear operation ?

Thanks for the links, thats a gold mine of information.

There are quite a few oscillator designs that include AGC mechanisms to keep a fairly constant amplitude and avoid much distortion. But I suggest increasing the emitter resistor a bit more. You may need to split it into two resistors, one with a shunt capacitor to provide quick starting, the second resistor not bypassed to provide low distortion after the quick start.

 

There are quite a few oscillator designs that include AGC mechanisms to keep a fairly constant amplitude and avoid much distortion. But I suggest increasing the emitter resistor a bit more. You may need to split it into two resistors, one with a shunt capacitor to provide quick starting, the second resistor not bypassed to provide low distortion after the quick start.

Increasing the emitter resistor over 137 ohm decays the oscillations. I tried 140 ohm.
The circuits seems to heavily rely on mathematics.

Even with the bypass capacitor approach, if its low impedance - ac signal will have high loop gain which will eventually cause
clipping/distortion.

 

Increasing the emitter resistor over 137 ohm decays the oscillations. I tried 140 ohm.
The circuits seems to heavily rely on mathematics.

Even with the bypass capacitor approach, if its low impedance - ac signal will have high loop gain which will eventually cause
clipping/distortion.

The purpose of the unbypassed portion is to set the DC bias, which normally affects the gain. The other place to adjust the AC feedback in a Colpits oscillator is the ratio of the two capacitors across the resonant portion. They were not equal in the original design, but lazy folks make them equal. In the original design article that I saw there was a formula that set the values so that the oscillator had only just enough gain. But that required some math, and so others started making them equal, and then filtering out the harmonics in the next few stages.